Validity of Hospital Discharge Diagnosis Codes for Hypertensive Disorders of Pregnancy

Kesselring, Cailey G.

Full Metadata

Overview

Year:: 2020
Contributor:: Kesselring, Cailey G. (creator); Savitz, David (Advisor); Danilack, Valery (Reader); Brown University. Department of Epidemiology (sponsor)
Genre:: theses
Subject:: Epidemiology; Maternal health services; Gestational Hypertension
Extent:: 6, 17 p.

Files

Description

Abstract:: Background Hypertensive disorders during pregnancy are prevalent worldwide. While research has been done to investigate the accuracy of ICD-9 coding, few studies have validated the use of ICD-9 codes in obstetric care and even fewer have conducted a detailed assessment of the different types of hypertensive disorders and codes. Objectives We sought to investigate the accuracy of hypertensive ICD-9 codes by comparing data from medical records, Rhode Island Department of Health birth certificate data, and ICD-9 codes from hospital administrative discharge data. Methods Data were collected as part of the Iatrogenic Delivery Study, a single-site retrospective cohort study of prenatal care and labor and delivery of pregnant women with a hypertensive disorder, diabetic disorder, or fetal growth restriction. Eligible women delivered a singleton live birth and had at least one of the previous conditions. Chart abstractions and birth certificate data were gathered for eligible women to gain more insight on maternal and neonatal outcomes of women. We validated hypertension diagnoses by calculating the proportion of individuals with a diagnosis code that met the criteria for the respective hypertensive disorder. With chart abstractions serving as gold standards, we calculated the positive predictive value (PPV) and negative predictive value (NPV) for ICD hospital administrative discharge ICD-9 codes and birth certificate data. To estimate the proportion of false negatives based on ICD-9 discharge diagnoses, we determined the proportion of individuals who qualified for the study for a non-hypertensive reason (i.e., fetal growth restriction or diabetes mellitus) and lacked an ICD-9 code indicative of hypertensive disorders but were found to qualify for a hypertension diagnosis upon chart review or indicated by the birth certificate data. Results We found a PPV of 88 percent [95% CI: 87, 89] and an NPV of 93 percent [92, 93] when examining discharge ICD-9 codes for any hypertensive disorder relative to a gold standard of chart abstracted hypertensive diagnoses. When we compared ICD-9 codes to chart abstractions for chronic hypertension, gestational hypertension, preeclampsia, and eclampsia diagnoses separately, we found a PPV of 99, 89, 68, and 33, respectively. When comparing birth certificate data to chart abstractions as the gold standard, we found a PPV of 90 and 96 for chronic hypertension and gestational hypertension, respectively. Lastly, when comparing birth certificate data to ICD-9 diagnostic codes as the gold standard, we found a PPV of 48 for chronic hypertension and 46 for gestational hypertension. Conclusion There is a wide variability when comparing hospital administrative discharge ICD-9 codes and birth certificate data to chart abstractions as the gold standard, with birth certificates notably less accurate than discharge diagnosis codes. For discharge diagnoses, the PPVs for chronic hypertension and gestational hypertension were quite good, and those for preeclampsia and eclampsia much weaker.
Notes:: Thesis (Sc. M.)--Brown University, 2020

Content

Access Conditions

Rights: In Copyright
Restrictions on Use: Collection is open for research.

Citation

Kesselring, Cailey G., "Validity of Hospital Discharge Diagnosis Codes for Hypertensive Disorders of Pregnancy" (2020). Epidemiology Theses and Dissertations. Brown Digital Repository. Brown University Library. https://repository.library.brown.edu/studio/item/bdr:1129366/

Relations

Collection:

Epidemiology Theses and Dissertations

Theses and Dissertations for the Epidemiology department.
...